Hi, I am having lots of failures generating extracts from my csv files. Not sure why this is happening. The failures occur both in Desktop (v2018.3 on Windows) as well as Server (v2018.2 Linux) using the following setup.
My setup is based on the directions here: Union Your Data - Tableau
- I have a data source which is a union of CSV files
- All CSV files in question end with spotDF.csv
- The union is setup to use a wildcard search, to include the pattern '*spotDF.csv'
- The union is setup to expand the search to sub-folders of the union's root file
On the filesystem, my data is synced in stages. First the required CSV files in question get copied to the folders to be used in the above union. Other associated files (not required, used, or referenced by Tableau) are copied next - those files (*.png, *.tiff, *.pdf, etc.) are co-located with the csv files that are needed by the extract.
When my extract gets kicked off, if ALL the files are already present in the sub-folders (CSVs, image files, PDFs, etc.), the extract succeeds. HOWEVER, if the extract starts while the NON-CSV files are still copying, the extract fails. I get an error such as:
com.tableausoftware.nativeapi.exceptions.DataSourceException: IsDirectory: Cannot get status of non-existent path '/mnt/MX_IMAGES/Microorganisms/2019-02/20190222/20190222__AST-2week_Study/.2019-02-23_09-08-09.tiff.IIsWSB'
or
com.tableausoftware.nativeapi.exceptions.DataSourceException: IsDirectory: Cannot get status of non-existent path '/mnt/MX_IMAGES/Microorganisms/2019-02/20190222/20190222__AST-2week_Study/AnnotatedImages/2019-02-22-11-04-00__Spots.png'
There is no question that the folders referenced above do exist.
The CSV file that is part of the union specified by the datasource exists here: '/mnt/MX_IMAGES/Microorganisms/2019-02/20190222/20190222__AST-2week_Study/.2019-02-23_spotDF.csv'
The file being complained about '2019-02-23_09-08-09.tiff.IIsWSB' is a temporary TIFF image file that was in the process of being copied into folder alongside the CSV file the extract needs. The TIFF file has NOTHING to do with the extract.
Similarly, the other file that caused an error on a different occasion is a transient png file that was there when the extract started and got removed by another process after the extract had commenced. But it's a PNG file, not a CSV file.
So why is Tableau having this hiccup as a result of files that have nothing to do with the union of CSVs? Recall, my pattern is "*spotDF.csv". The files Tableau complained about do not fit that pattern.
I hope someone can shed some light on this or provide a workaround. I cannot hold off on starting the extract until all the data in the subfolders has been synced. The CSV files, which are the basis for the data source and the extract, are present and accounted for. Shouldn't that be all that matters?
Thanks,
E
Hi Elias, my guess is that when Tableau is scanning your filesystem and it encounters an error, it gives up. It doesn't know the cause of the error, and so it assumes it could be something real. if you want an official answer, your best bet is to contact technical support; perhaps this is something that was never considered.
A possible workaround is to store the non-CSV files elsewhere, and when the copying is complete, move the files into place or create symbolic links (both of which are quick operations. (That's assuming you really need the extra files to be in the same location.)
